In a regional and industry-first, du, BubbleRAN, and Khalifa University have demonstrated how Arabic voice commands can operate autonomous 5G networks — turning spoken “intents” into verified network actions using O-RAN and AI-RAN compliant Telco APIs.
This breakthrough combines reasoning-capable Arabic-native telecom LLMs with deep domain knowledge such as regulations, optimization logic, and user behavior. Engineers can now manage networks using Arabic voice, accelerating automation, reducing human error, and ensuring full cultural and linguistic accuracy without English translation.
du CTO Saleem AlBlooshi said the innovation will make private 5G more efficient and scalable, while BubbleRAN CEO Prof. Navid Nikaein called it “the world’s first step toward autonomous 5G/6G.” Khalifa University’s Prof. Merouane Debbah noted that AI agents, TelecomGPT Arabic, and rApps played a key role in enabling zero-touch network automation.
The system uses an Arabic-native telecom LLM that understands dialects and telecom jargon, reasons step-by-step, and validates decisions before applying them via rApps. BubbleRAN’s MX-AI platform orchestrates multiple AI agents to automate complex lifecycle operations.
BubbleRAN’s MX-AI solution (Crimson release) is now available for deployment. The base Arabic LLM used in this demo is public on Hugging Face for non-commercial research.
